Project Abstract In patients with vascular disease, clear delineation of the arteries of interest is essential for diagnosis and treatment planning. While a number of angiography techniques are available nowadays, non- contrast-enhanced (NCE) magnetic resonance angiography (MRA) is considered a highly promising modality due to non-invasiveness as opposed to digital subtraction angiography (DSA), and the absence of side effects unlike computerized tomography (involving ionizing radiation) and gadolinium- contrast-enhanced MRA (involving the risk of nephrogenic systemic fibrosis). As such, many NCE MRA methods have been explored in the past decades, yet are still at some distance from solid clinical acceptance due to issues such as limited angiographic coverage and spatial resolution, and long scan time (particularly when multiple acquisitions are needed). This project will focus on novel NCE MRA methods which achieve high vessel contrast, high spatial resolution in all 3 dimensions and large angiographic coverage in short scan time. At the core of the proposed method is velocity-selective (VS) magnetization preparation which generates angiographic contrast by suppressing background materials while preserving arterial blood based on their velocity. Due to spatially non-selective nature, VS preparation can be combined with 3D encoding with large FOV and high spatial resolution in all three dimensions unlike inflow-based approaches such as time- of-flight. Also, VS preparation generates positive angiographic contrast directly from single acquisition, as opposed to existing flow-sensitive approaches that require two acquisitions to be subtracted. We will first develop VS excitation pulse sequences which are robust to potential variation in B0 and B1 field as well as eddy current of VS gradient. Areas of investigation will include incorporation of multiple refocusing pulses, Malcom-Levitt phase cycling, and two successive VS preparations. Exploiting the sparsity of background-suppressed image of VS-MRA, we will achieve high-rate scan acceleration by combining compressed sensing with parallel imaging reconstruction. While the proposed VS-MRA is applicable for diverse vascular territories, this project will focus on peripheral angiography and cerebral angiography as representative clinical applications. For each of the two applications, we will optimize VS-MRA pulse sequences by selecting optimal number of refocusing pulses and type of refocusing composite pulse based on estimates of B0 and B1 offsets. Using the optimized protocol, lastly, diagnostic performance will be evaluated in patients with x-ray angiography as the reference.